By Alex Veytsel – One of the panels I most regret missing at Data Center Dynamics annual London show last week, where our CEO spoke was a discussion of the future of data centers in 10-20 years. Fortunately I got to participate in the prep session for it in the speaker’s lounge, where brilliant folks like Ian Bitterlin, David Gauthier, and Ed Ansett kicked around ideas ranging from the pragmatic to the futuristic under the steady guiding hand of Ambrose McNevin (and I hope I got a shout-out if there was mention of pendulums or whale oil). Which got me to thinking – whether the amount of change is great or small, how will the way services are bought change over that time? What will be the sourcing process in 10 years?
The short answer: buyers will demand more transparency; suppliers will give them just enough to continue finding more margin and feed the sales machine; and sourcing advisors will build a toolset to fit the right service for your business more automatically, with more just-in-time adjustment.
First, I would say that the commercial aspects of outsourcing are still the most stubbornly un-changeable ones. Just this year, I have reviewed a still-active contract written on Exodus Communications paper and multiple invoices based on racks and square feet in an age where everyone thinks of watts consumed as the primary / sole metric.
And 12 years in, RampRate’s mission has not changed much – there’s still billions in wasted IT spend through bad decisions; still too many people listening to salespeople instead of seeking out data, and a great terror of being exposed for one’s incompetence that leads buyers to run from their best sources of improvement.
That said, some things are starting to move. Cloud services, especially IaaS, brought unprecedented transparency to pricing. Not that there isn’t an opportunity to get off the price list if you’re big enough (or smart enough), or lack of clarity in the many customizations, or a pretty sorry state of tying pay to performance… well, perhaps, it’s not that big of a sea change – somewhat mirroring how much of the good old managed hosting paradigm still lives beneath the moniker of cloud. But it’s a start.
However, it’s clear that a buyer in 10-20 years will clearly want and expect more visibility into what they’re buying.
This will manifest in several ways:
- Detailed price lists that are linked to the right pricing metric (no more mysterious “20 racks [10 useable]” line items that we see today)
- Public or well-known scale discounts
- Activation of scale discounts / surcharges based on demand (you use more, you get into the higher bracket)
- At minimum a gold / silver / platinum service level menu (or better yet, a continuum of options) that discounts services based on how much latency, downtime, and performance consistency a buyer expects
- Closer approach to pricing parity / true commoditization due to easier switchover / migration
So does that mean that I think RampRate will be out of business? Well, hardly. Even as much of the pressure of the competitive process (and the embedding of sourcing advisors into buying departments everywhere) will drive some improvements, others will stubbornly resist change:
- Direct sales forces will still retain immense influence, controlling introductory offers / strategic discounts and monopolizing certain large accounts, requiring continued fight for channel integration and reduction in commissionable layers
- Units of purchase will still be influenced by marketing, which will actively seek to differentiate their firm’s services by selling their own version of compute units that’s harder to match up to competitors in an “apples-to-apples” comparison
- Buying inefficiencies will still be encouraged by “all you can eat” pre-payment plans that offer the prospect of a small discount if things go exactly according to plan, but trap the buyer in an unworkable deal if business requirements change.
Compared to holographic storage, fusion power, and spending 80% of the world’s energy on computing, this seems like an awful lack of imagination. But considering that we see a predatory business logic still working even after the singularity happens and we’re all cyborgs built by nanomachines, perhaps it should not be surprising that the 10 year horizon is closer than you’d think. But that said, there will be better weapons on the buyer side in the arms race of data vs. salesmanship. We know because we’re building them already:
- Tools to optimize allocation of services (by application, business unit, or even process) to the right amount of fault tolerance (and therefore the right price) within each supplier, and across multiple suppliers concurrently. Just as a single full-IT outsource contract is rapidly becoming a dinosaur, it will become more rare to be attached to a single data center, a single network, a single cloud, etc.
- An electronic due diligence process that tracks real-world performance across hundreds of buyers to give solid data on actual uptime, latency, PUE, and other metrics, rather than taking supplier’s word
- A translation engine that not only cuts through the differentiation efforts of suppliers to the underlying commodity metrics, but also allows you to choose from data center, managed, or cloud depending on where on the maturity curve your IT shop is and how strategic IT really is for you.
- An automatic contract writer that inserts the right terms and service levels for your business’ needs and finds the appropriate service within the supplier portfolio to support them.
This improvement in transparency will help drive redirection of some of the supplier differentiation budget from sales and marketing to transparency and service, lifting the tide for all boats – even non-users of sourcing advisory toolsets. It will also speed up the re-contracting cycle to where deals of a few months will be the norm even outside of the cloud environment – with a smoother, more well-oiled supplier selection process, a renewal / mid-term adjustment will be a less arduous process. And maybe, just maybe, the rest of the world’s outsourcing failure rates will start to resemble those of RampRate’s customers today (<1% for those keeping track).